Comments on: group think

By: Steve

Steve — Fri, 18 Apr 2008 19:09:27 +0000

Great post. I have to comment on one aspect of it: any description or analogy to the WTO protests in Seattle is probably wildly off the mark. Are you comparing to what _actually happened_ or what was reported on TV? Because the two were wildly different.

I watched it on TV from my office on the outskirts of downtown and thought my city was in riot mode. After two hours strolling across all of downtown and the “riot,” it was evident to me that was really happening: TV crews making hay while the sun was shining. There was a legitimate protest, but it was far scarier on television than in person.

The research being done is valuable and worth pursuing, but WTO Seattle highlights some of the challenges moving from theory to reality.

By: fruminator

fruminator — Mon, 14 Apr 2008 03:48:31 +0000

I don’t know Steve, was Black Scholes designed to _describe_ the way they thought options _were_ priced, or to _prescribe_ the way they thought they _should_ be priced? If it’s the former, then you are spot on. If it’s the latter, then perhaps not so much.

[Editor: From my reading, it was absolutely designed to be descriptive, not prescriptive. s.

By: rab

rab — Tue, 08 Apr 2008 18:41:31 +0000

The staggering increase in surveillance technologies and their implementation by military, police, government and business follow the old pattern of - if we have a capability, have paid for it, and can deploy it, then we will test it and use it. The human being or individual is no match for the models imposed on them. Once deployed such practices multiply, and resistive practices require “groups” to be effective. Infiltration is the common response to limit or redirect such groups. The question I would like to pose is - What kind of countermeasures are available to those whose “crowd” wishes to remain either anonymous, or off the grid? It is here that participation in social protest or unregulated activity must multiply in its diversity and deployment, outwitting, like the border-crossers from Mexico, every lame fence or surveillance grid applied to control them.

By: sgs

sgs — Tue, 08 Apr 2008 02:42:10 +0000

Biologist — thanks for your comments. I definitely deserve some grief for the perfunctory hand-waving that serves as my conclusion. (What can I say? It was already at 4,500 words and I just hadda bring it home.) That said, I’m not sure your convincing evisceration of several straw men really moves the ball forward.

First, you point out that simplification, and reductionist models, are the bulwark of modern science. This does not come as news to me. Indeed, I was careful to emphasize that I was not impugning the fundamental efficacy of the approach. Instead, my concern was with cultural impact– the moral angle.

Then you ask why the cultural affects will be any more deleterious than those from, say, drug-development research, or physics. And you wonder what the heck my phrase “cheapening ourselves” actually means.

The second question is on-target — more of my handwaving — but the first is just rhetorical bunting. As Marvin Minsky famously told Steward Brand, “Don’t get blinded with analogies and similarities– always look for the _differences_.” The cultural affects of using simplified models of human behavior will be different for epidemiology, because what’s being studied is different.

But, enough sparring– let me just answer your main, underlying question: what makes crowd theory so “pernicious”? Based on people’s comments, a surprising number of readers quickly tumbled to the answer: My biggest concern is that the models will be self-fulfilling.

I mentioned Black-Scholes in the essay, in part because it is such a wonderful example of this phenomena. As Donald MacKenzie explains in his brilliant book “An Engine, Not a Camera”, option prices before 1973 show very little agreement with the Black-Scholes predictions. But after the equations were published in 1974, the divergence slowly shrank away.

Yes, Black-Scholes became an accurate model of option pricing because… people began using the formula to price options. But it was self-fulfilling in a deeper sense, as well. Just as formulas in physics make simplifying assumptions – a frictionless incline, no wind resistance — Black-Scholes assumed zero transaction costs, unlimited borrowing at the riskless interest rate, and unconstrained short-selling. These were all wildly unrealistic in 1973. But as regulators adopted Black-Scholes to govern everything from risk at financial institutions to executive compensation, the model’s assumptions rode along like stowaways, becoming deeply embedded in economic policy. Those “unrealistic assumptions” are now empirical fact. To a very real extent, the world was remade to fit the model, instead of the model being tweaked to fit the world.

This, then, is why crowd theory threatens to “cheapen ourselves”. Because it tries to explain human behavior, it will also impact human behavior in a way that, say, biological models never could. I fear the impact will be that, in subtle and seemingly innocuous steps, we will begin to act more like automatas, and less like humans.

-sgs.

By: dren

dren — Tue, 08 Apr 2008 00:23:10 +0000

People can’t even figure out horse racing. What chance do they have of figuring out each other?

I don’t know whether or not to be scared by this technology. Certainly, if anyone wanted to track me, it would be incredibly feasible for them to do so. It would also be extremely easy to accurately predict my behavior on a level of how late I stay up each night, when I go to bed, what route I take to work, etc. I don’t know what is so scary about that. I have no reason to believe Verizon isn’t doing it right now. They have a GPS in my phone. For all I know it could stay on even if I pull my phone’s battery out. Verizon has or could easily have all the data on me. They probably don’t even need a computer, just someone to read my file for a couple hours, in order to figure out my daily patterns. Perhaps I should be more frightened by this but I just don’t see how it’s possible to truly be anonymous or private.

I recall a recent news piece I saw about a Comcast exec saying he’d like it if the cable box had a camera in it to identify which members of the household were watching TV for the purpose of tailoring ad content. Naturally, people don’t want a camera in their home showing what they’re doing. But if Comcast were to collude with Verizon to see when I was at home, they’d be able to serve up much improved ad content. Not that Comcast and Verizon would ever team up since cable and fios are competing services, but you get the idea.

By: jaguar19

jaguar19 — Mon, 07 Apr 2008 18:12:32 +0000

The arguments made here about the inherently evil nature of studies of group behavior miss the point. Just as crowd behavior at a demonstration can be modeled, are you not also modeling the behavior of the police that they are interacting with? Indeed, could not this very kind of modeling be used to determine/influence the behavior of the police? I would think that military thinking (sic) is probably more easily modeled as it is in many ways so goverened by simple rules of engagement. Or consider the mentality of the overwhelming amount of governmental bureaucracy. Is there any more beehive-like behavior than that of a department of government bureaucrats?

By: Sur Surac

Sur Surac — Mon, 07 Apr 2008 15:16:16 +0000

If the technology doesn’t lead us to be controlled or less humans, we still take the risk of self-fulfill the profecy

By: That college guy

That college guy — Sun, 06 Apr 2008 23:49:20 +0000

Firstly, I would like to focus on the fact that the crowd theory is still based on the predictability of crowds, this being a crowd that has a singular vision will not divert from its course unless there is an external. This external being a; non-threatening, predictable, entity. the mention of the mathematics of how gas moves, related then to how there are similarities made me think of the fact that as mentioned at quantum level particles still react unpredictably, and that there is still no mathematical way to predict how gas will move in a room, because there are too many variables.
Crowd theory suffers the same problem. The empirical is done in controlled environments, and can thus never forecast or truly predict, all it does is force empirical information on a situation and hopes that there will be no unknown.
Here I would like to mention Nassim Nicholas Talebs book The Black Swan (READ IT!), and the Soweto uprisings of the 1980’s (south Africa) and most other protests that turn bad; the protest may be controlled by an ‘overseer’ but this higher will never be able to control the base events that cause a violent protest, that being in many circumstances, a random shot from the police that acts like a catalysts.
Secondly on the paranoia of the commentators; there will always be a struggle between the ‘Man’ and the individual, it is this check that balances what can and can’t be done. This might sound naive, I know, but the reality is that using information such as crowd theory will only streamline society and allows for increased production and speed up the rate of development. We have moved beyond an Orwell like 1984, where government needs to control the individual. The individual bought himself into such a world when we accepted capitalism.
Now, for arguments sake, I would like to agree by saying crowd theory will change our world, however, the impact of crowd theory will only be visible in a society where the individual agrees to follow the suggestions and will never be a tool that has the power to control the individual. The true benefactor of such information will be the business sector, for they will have the power to control large groups, employers, without facing resistance.
Laters, peace.

By: Brendan

Brendan — Sun, 06 Apr 2008 21:32:24 +0000

It seems little concerning that any group who has power to influence people and relies on these models will have an incentive to coerce people into thinking like their model suggests.

Since it is unlikely that models will any time soon/ever involve the kind of complexity of actual society, it seems that widespread introduction of these models will likely result in more pressure towards conformity.

By: a johnstone

a johnstone — Sun, 06 Apr 2008 19:37:35 +0000

Excellent post, very elucidating. I think people often forget their behaviors are essentially controlled by string theory. Technology is bringing us closer everyday to a parallel processing of our world. A scary thought that someone could know what was going happen hours or days before it did.

It’s wonderful to find thought provoking content like this online, keep up the great work.

By: mike s

mike s — Sun, 06 Apr 2008 15:14:21 +0000

“It is important for the common good to foster individuality; for only the individual can produce the new ideas which the community needs for its continuous improvement and requirements - indeed, to avoid sterility and petrification.”

-Albert Einstein

By: AG

AG — Sun, 06 Apr 2008 12:23:42 +0000

I applaud (and to some degree share) your spirit of resistance, casey, but I promise you that willful and contrarian and even random behavior are as easy to model as lockstep conformity in scenarios such as these.

By: Pernicious chnit

Pernicious chnit — Sat, 05 Apr 2008 23:18:51 +0000

Congratulations to Biologist for using the work “pernicious” with a frequency higher than any other reply or feedback in the history of the Internet or electronic bulletin boards.

By: casey

casey — Sat, 05 Apr 2008 22:42:19 +0000

This article has inspired me to take paths less taken, and in general behave as unusually and randomly as I can get away with. Not only does it help you feel more alive, it keeps Them from being able to create a model of me.

Are you alive, or a model?

By: kedabra

kedabra — Sat, 05 Apr 2008 21:38:07 +0000

the problem that strikes me with this kind of technology, and also with data trawling technologies, is that it is easier and (crucially) cheaper to build simplified models of human behaviour , our interactions and patterns of behaviour , and then detect suspicious abberations for investigation, than it is to detect and investigate instances of actual criminal behaviour.
basically we could end up defining normal behaviour, and harassing those who do not fit the model, instead of pursuing those who have actually committed crimes. this will pressure people to become more conformist , less individual, and not to question authority, which will lead to stagnation and degradation of our culture and society.